Plant species – The selected model species included Noccaea caerulescens (Brassicaceae) targeting zinc, Biscutella laevigata (Brassicaceae) and Silene latifolia (Caryophyllaceae) targeting thallium, Phytolacca octandra (Phytolaccaceae) targeting manganese, and Pityrogramma calomelanos (Pteridaceae) targeting arsenic. Furthermore, we employed the well-known metal tolerant species capable of producing root exudates, Lupinus albus (Fabaceae). The intensively studied hyperaccumulator N. caerulescens from Europe has the ability to hyperaccumulate Ni, Zn, Pb and Cd with accessions differing in their ability to tolerate and accumulate (Assunção et al. 2003; Deng et al. 2014; Gonneau et al. 2014; Schwartz et al. 2003). The level of bioconcentration in N. caerulescens is extreme as it can accumulate up to 52,000 µg g-1 foliar Zn (Zhao et al. 2003), whilst it can also attain 3410 µg g-1 foliar Cd (Reeves et al. 2001, 2018a,b). Biscutella laevigata can attain > 3 wt% foliar Tl when growing on soils with ~ 400 µg Tl g-1 (Pošćić et al. 2015) and S. latifolia up to 1500 µg Tl g-1 (Escarré et al. 2011). Pityrogramma calomelanos is an As hyperaccumulator that can reach up to 8350 µg As g− 1 in its fronds (Francesconi et al. 2002). Phytolacca octandra and L. albus produce root exudates, thereby increasing phytoavailable metal concentrations (Dinkelaker et al. 1989; Lambers et al. 2015). Moreover, Phytolacca species are Mn hyperaccumulators and may be beneficial in reducing Mn-toxicity to co-cultivated plants via the removal of available Mn from the rhizosphere (Pollard et al. 2009). Phytoavailability of the target metal(s) in the soil or substrate is a key consideration for the effectiveness of phytoextraction (Nkrumah et al. 2016, 2021) and we hypothesize that co-cultivating root exudates-producing plants with selected model hyperaccumulator plants could increase metal accumulation, especially in relatively low phytoavailable metal substrates. Finally, as a legume, L. albus is a nitrogen-fixing species which may improve the fertility of the substrate for the co-cultivated hyperaccumulator plants.
Experimental design – The seed accession of N. caerulescens (Navacelles, France, non-metalliferous) was chosen purposely because it has a particularly high bioconcentration factor for Zn (Escarré et al. 2000). The seed accessions of B. laevigata and S. latifolia originated from Saint-Laurent-le-Minier, Southern France (metalliferous). Seeds of P. octandra were collected near Brisbane, Australia, while seeds of L. albus were purchased from a commercial supplier in Brisbane and Pityrogramma calomelanos sporelings were collected from north Queensland. The experiments were conducted in plastic boxes (45 × 30 × 12 cm, L, W, H) in which a 1 cm layer of plastic granules (for irrigation) was covered by a plastic mesh and then filled with tailings mixes consisting of 90:5:5 and 75:10:15 (w/w/w) tailings/peat moss/perlite for Dugald River Mine and Mt Isa Mine tailings, respectively. The Dugald River Mine tailings have chemical properties that are generally amenable to plant growth (pH 6.75, relatively low salinity), although it has high concentrations of Mn (0.41 wt%), Zn (22,000 mg kg-1), Cd (50 mg kg-1), Tl (30 mg kg-1), Pb (8220 mg kg-1). The Mt Isa Mine tailings had relatively lower Mn (0.17 wt%), Zn (10,100 mg kg-1), Cd (35 mg kg-1), Pb (6240 mg kg-1), but higher Tl (75 mg kg-1). Seeds of N. caerulescens, B. laevigata and S. latifolia were germinated separately on Gelzan gel in 2 mL Eppendorf tubes (made from 0.5-strength Hoagland’s solution). The seeds were then vernalized for one week at 3°C and acclimatized for four days at 26°C (van der Zee et al. 2021). Seeds of P. octandra were treated with sulphuric acid for 15 mins and rinsed with ultrapure water before sowing in a 1:1 mix of perlite and vermiculite. The seedlings (approx. 1 cm in size) were transplanted to the boxes. Four treatments with 16 biological replicates were undertaken on the Dugald River Mine substrate: B. laevigata only, B. laevigata + N. caerulescens, B. laevigata + L. albus and N. caerulescens only. In addition, seven treatments with 16 biological replicates were undertaken on the Mt Isa Mine substrate: B. laevigata only, B. laevigata + P. calomelanos, B. laevigata + L. albus, N. caerulescens only, B. laevigata + P. octandra, S. latifolia only and S. latifolia + P. octandra. The plants were grown for a total period of 12 weeks, harvested and then separated into root and shoot fractions for analysis as described below.
Bulk analysis of plant tissue samples – All of the plant material samples were thoroughly washed with demineralised water and then oven dried at 70°C for 3 days. The samples were weighed and then ground to a fine powder in an impact mill at 15,000 rpm (IKA Tube Mill 100 control with disposable titanium blades) and then 100 ± 5 mg of each sample was weighed into 6 mL polypropylene tubes. The samples were pre-digested using 2 mL HNO3 (70%) for 24 hr before digestion in a block heater (Thermo Scientific™ digital dry bath) for a 2-hr programme (1 hr at 70°C followed by 1 hr at 125°C). The digestates were then brought to volume (10 mL) with ultrapure water before analysis with Inductively coupled plasma atomic emission spectroscopy (ICP-AES) with a Thermo Scientific iCAP 7400 instrument for macro-elements (Mg, P, S, K, Ca) and trace-elements (Mn, Fe, Cu, Zn, As, Cd, Tl, Pb) in radial and axial modes depending on the element and expected analyte concentration. All elements were calibrated with a 4-point curve covering analyte ranges in the samples.
Collection and analysis of tailings samples – Samples of the respective substrates collected from 5–20 cm depth were air-dried and sieved through a 630 µm screen. The pH was measured in a 1:2.5 substrate to water slurry after 2 hr mixing on an end-over-end shaker and a 1-hr rest. Sub-samples were weighed (100 ± 5 mg) in quartz digestion vessels and 5 mL HNO3 (70%) and 2 mL HCL (37%) were added. The samples were then digested for 15 min at 80% power using a ColdBlock system (CB15S 15 channel system, ColdBlock Technologies Inc) which uses high-intensity infrared irradiation to aid rapid acid digestion (Wang et al. 2014). The digestates were quantitatively transferred to 50 mL tubes, brought to volume (40 mL) and filtered (Whatman® Grade 1 filter paper) before analysis with ICP-AES.
micro-X-ray fluorescence elemental mapping – The UQ micro-XRF facility is a custom-built system manufactured by IXRF which consists of two 50 kV sources (1000 µA) fitted with polycapillary focussing optics: XOS microfocus Mo-target tube producing 17.4 keV X-rays (flux of 2.2 × 108 ph s-1) focussing to 25 µm and a Rh-target tube producing 20.2 keV (flux of 1.0 × 107 ph s-1) focussing to 5 µm. The system is fitted with two silicon drift detectors (SDD) of 150 mm2 coupled to a XIA Mercury X4 signal processing unit. The fresh foliar samples were mounted between two sheets of 4 µm Ultralene thin film in a tight sandwich to limit evaporation and analysed within 10 minutes after excision. The mounted samples between Ultralene thin film were stretched over a Perspex frame magnetically attached to the x-y motion stage at atmospheric temperature (~ 20°C). The UQ microXRF facility acquired the XRF spectra in mapping mode using the instrument control package, Iridium (IXRF systems), from the sum of counts at the position of the principal peak for each element. They were then exported into ImageJ as greyscale 8-bit TIFF files, internally normalised such that each image covered the full dynamic range and displayed using ImageJ’s “Fire” lookup table.
Statistical analyses
Statistical analyses were performed using OriginPro 2021 (https://www.originlab.com/). The elemental concentrations of the plant fractions were presented in Tables and plotted in Figures as mean ± standard error. Significant differences were determined by ANOVA, separated by Tukey’s honestly significant difference (HSD) test (p < 0.05) and indicated by different letters.